From: "Steve Wise" <swise@opengridcomputing.com>
To: "'Christoph Hellwig'" <hch@infradead.org>
Cc: <dledford@redhat.com>, <sagig@mellanox.com>,
<ogerlitz@mellanox.com>, <roid@mellanox.com>,
<linux-rdma@vger.kernel.org>, <eli@mellanox.com>,
<target-devel@vger.kernel.org>, <linux-nfs@vger.kernel.org>,
<trond.myklebust@primarydata.com>, <bfields@fieldses.org>
Subject: RE: [PATCH V3 1/5] RDMA/core: Transport-independent access flags
Date: Mon, 6 Jul 2015 09:23:42 -0500 [thread overview]
Message-ID: <000c01d0b7f7$5c5f02c0$151d0840$@opengridcomputing.com> (raw)
In-Reply-To: <20150706052515.GA1109@infradead.org>
> -----Original Message-----
> From: Christoph Hellwig [mailto:hch@infradead.org]
> Sent: Monday, July 06, 2015 12:25 AM
> To: Steve Wise
> Cc: dledford@redhat.com; sagig@mellanox.com; ogerlitz@mellanox.com; roid@mellanox.com; linux-rdma@vger.kernel.org;
> eli@mellanox.com; target-devel@vger.kernel.org; linux-nfs@vger.kernel.org; trond.myklebust@primarydata.com; bfields@fieldses.org
> Subject: Re: [PATCH V3 1/5] RDMA/core: Transport-independent access flags
>
> On Sun, Jul 05, 2015 at 06:22:00PM -0500, Steve Wise wrote:
> > The semantics for MR access flags are not consistent across RDMA
> > protocols. So rather than have applications try and glean what they
> > need, have them pass in the intended roles and attributes for the MR to
> > be allocated and let the RDMA core select the appropriate access flags
> > given the roles, attributes, and device capabilities.
> >
> > We introduce rdma_mr_roles and rdma_mr_attributes that enumerate the
> > possible roles and attributes for a MR. These are documented in the
> > enums themselves.
> >
> > New services exported:
> >
> > rdma_device_access_flags() - given the intended MR roles and attributes
> > passed in, return the ib_access_flags bits for the device.
> >
> > rdma_get_dma_mr() - allocate a dma mr using the applications intended
> > MR roles and MR attributes. This uses rdma_device_access_flags().
> >
> > rdma_fast_reg_access_flags() - return the ib_access_flags bits needed
> > for a fast register WR given the applications intended MR roles and
> > MR attributes. This uses rdma_device_access_flags().
> >
> > Signed-off-by: Steve Wise <swise@opengridcomputing.com>
> > ---
> >
> > drivers/infiniband/core/verbs.c | 30 +++++++++++
> > include/rdma/ib_verbs.h | 106 +++++++++++++++++++++++++++++++++++++++
> > 2 files changed, 136 insertions(+), 0 deletions(-)
> >
> > diff --git a/drivers/infiniband/core/verbs.c b/drivers/infiniband/core/verbs.c
> > index bac3fb4..f42595c 100644
> > --- a/drivers/infiniband/core/verbs.c
> > +++ b/drivers/infiniband/core/verbs.c
> > @@ -1144,6 +1144,36 @@ struct ib_mr *ib_get_dma_mr(struct ib_pd *pd, int mr_access_flags)
> > }
> > EXPORT_SYMBOL(ib_get_dma_mr);
> >
> > +int rdma_device_access_flags(struct ib_pd *pd, int roles, int attrs)
> > +{
> > + int access_flags = attrs;
>
> Please add an assert for the values that are allowed for attrs.
>
> It also would be highly useful to add a kerneldoc comment describing
> the function and the parameters. Also __bitwise sparse tricks
> to ensure the right flags are passed wouldn't be a too bad idea.
>
Can you explain the "__bitwise sparse tricks"? Or point me to an example.
> > +/**
> > + * rdma_mr_attributes - attributes for rdma memory regions
> > + */
> > +enum {
> > + RDMA_MRA_ZERO_BASED = IB_ZERO_BASED,
> > + RDMA_MRA_ACCESS_ON_DEMAND = IB_ACCESS_ON_DEMAND,
> > +};
>
> Why not specfiy these as separate nuerical namespace?
>
To avoid having to translate them.
> > +/**
> > + * rdma_device_access_flags - Returns the device-specific MR access flags.
> > + * @pd: The protection domain associated with the memory region.
> > + * @roles: The intended roles of the MR
> > + * @attrs: The desired attributes of the MR
> > + *
> > + * Use the intended roles from @roles along with @attrs and the device
> > + * capabilities to generate the needed access rights.
> > + *
> > + * Return: the ib_access_flags value needed to allocate the MR.
> > + */
> > +int rdma_device_access_flags(struct ib_pd *pd, int roles, int attrs);
>
> Oh, here we have kerneldoc comments, just in the wrong place. Please
> move them to the function defintion.
Ok.
WARNING: multiple messages have this Message-ID (diff)
From: "Steve Wise" <swise@opengridcomputing.com>
To: 'Christoph Hellwig' <hch@infradead.org>
Cc: dledford@redhat.com, sagig@mellanox.com, ogerlitz@mellanox.com,
roid@mellanox.com, linux-rdma@vger.kernel.org, eli@mellanox.com,
target-devel@vger.kernel.org, linux-nfs@vger.kernel.org,
trond.myklebust@primarydata.com, bfields@fieldses.org
Subject: RE: [PATCH V3 1/5] RDMA/core: Transport-independent access flags
Date: Mon, 6 Jul 2015 09:23:42 -0500 [thread overview]
Message-ID: <000c01d0b7f7$5c5f02c0$151d0840$@opengridcomputing.com> (raw)
In-Reply-To: <20150706052515.GA1109@infradead.org>
> -----Original Message-----
> From: Christoph Hellwig [mailto:hch@infradead.org]
> Sent: Monday, July 06, 2015 12:25 AM
> To: Steve Wise
> Cc: dledford@redhat.com; sagig@mellanox.com; ogerlitz@mellanox.com; roid@mellanox.com; linux-rdma@vger.kernel.org;
> eli@mellanox.com; target-devel@vger.kernel.org; linux-nfs@vger.kernel.org; trond.myklebust@primarydata.com; bfields@fieldses.org
> Subject: Re: [PATCH V3 1/5] RDMA/core: Transport-independent access flags
>
> On Sun, Jul 05, 2015 at 06:22:00PM -0500, Steve Wise wrote:
> > The semantics for MR access flags are not consistent across RDMA
> > protocols. So rather than have applications try and glean what they
> > need, have them pass in the intended roles and attributes for the MR to
> > be allocated and let the RDMA core select the appropriate access flags
> > given the roles, attributes, and device capabilities.
> >
> > We introduce rdma_mr_roles and rdma_mr_attributes that enumerate the
> > possible roles and attributes for a MR. These are documented in the
> > enums themselves.
> >
> > New services exported:
> >
> > rdma_device_access_flags() - given the intended MR roles and attributes
> > passed in, return the ib_access_flags bits for the device.
> >
> > rdma_get_dma_mr() - allocate a dma mr using the applications intended
> > MR roles and MR attributes. This uses rdma_device_access_flags().
> >
> > rdma_fast_reg_access_flags() - return the ib_access_flags bits needed
> > for a fast register WR given the applications intended MR roles and
> > MR attributes. This uses rdma_device_access_flags().
> >
> > Signed-off-by: Steve Wise <swise@opengridcomputing.com>
> > ---
> >
> > drivers/infiniband/core/verbs.c | 30 +++++++++++
> > include/rdma/ib_verbs.h | 106 +++++++++++++++++++++++++++++++++++++++
> > 2 files changed, 136 insertions(+), 0 deletions(-)
> >
> > diff --git a/drivers/infiniband/core/verbs.c b/drivers/infiniband/core/verbs.c
> > index bac3fb4..f42595c 100644
> > --- a/drivers/infiniband/core/verbs.c
> > +++ b/drivers/infiniband/core/verbs.c
> > @@ -1144,6 +1144,36 @@ struct ib_mr *ib_get_dma_mr(struct ib_pd *pd, int mr_access_flags)
> > }
> > EXPORT_SYMBOL(ib_get_dma_mr);
> >
> > +int rdma_device_access_flags(struct ib_pd *pd, int roles, int attrs)
> > +{
> > + int access_flags = attrs;
>
> Please add an assert for the values that are allowed for attrs.
>
> It also would be highly useful to add a kerneldoc comment describing
> the function and the parameters. Also __bitwise sparse tricks
> to ensure the right flags are passed wouldn't be a too bad idea.
>
Can you explain the "__bitwise sparse tricks"? Or point me to an example.
> > +/**
> > + * rdma_mr_attributes - attributes for rdma memory regions
> > + */
> > +enum {
> > + RDMA_MRA_ZERO_BASED = IB_ZERO_BASED,
> > + RDMA_MRA_ACCESS_ON_DEMAND = IB_ACCESS_ON_DEMAND,
> > +};
>
> Why not specfiy these as separate nuerical namespace?
>
To avoid having to translate them.
> > +/**
> > + * rdma_device_access_flags - Returns the device-specific MR access flags.
> > + * @pd: The protection domain associated with the memory region.
> > + * @roles: The intended roles of the MR
> > + * @attrs: The desired attributes of the MR
> > + *
> > + * Use the intended roles from @roles along with @attrs and the device
> > + * capabilities to generate the needed access rights.
> > + *
> > + * Return: the ib_access_flags value needed to allocate the MR.
> > + */
> > +int rdma_device_access_flags(struct ib_pd *pd, int roles, int attrs);
>
> Oh, here we have kerneldoc comments, just in the wrong place. Please
> move them to the function defintion.
Ok.
next prev parent reply other threads:[~2015-07-06 14:23 UTC|newest]
Thread overview: 223+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-05 23:21 [PATCH V3 0/5] Transport-independent MRs Steve Wise
2015-07-05 23:22 ` [PATCH V3 1/5] RDMA/core: Transport-independent access flags Steve Wise
2015-07-06 5:25 ` Christoph Hellwig
2015-07-06 14:23 ` Steve Wise [this message]
2015-07-06 14:23 ` Steve Wise
2015-07-07 8:58 ` 'Christoph Hellwig'
2015-07-06 7:09 ` Haggai Eran
2015-07-06 7:09 ` Haggai Eran
2015-07-06 14:29 ` Steve Wise
2015-07-06 14:29 ` Steve Wise
2015-07-07 14:17 ` Steve Wise
2015-07-07 14:17 ` Steve Wise
2015-07-07 14:34 ` Haggai Eran
2015-07-07 14:34 ` Haggai Eran
2015-07-07 14:46 ` Steve Wise
2015-07-07 14:46 ` Steve Wise
2015-07-07 15:07 ` Haggai Eran
2015-07-07 15:07 ` Haggai Eran
2015-07-06 7:53 ` Sagi Grimberg
2015-07-06 7:53 ` Sagi Grimberg
2015-07-06 14:37 ` Steve Wise
2015-07-06 14:37 ` Steve Wise
2015-07-06 16:17 ` Sagi Grimberg
2015-07-06 16:17 ` Sagi Grimberg
2015-07-06 16:55 ` Steve Wise
2015-07-06 16:55 ` Steve Wise
2015-07-07 9:00 ` Christoph Hellwig
2015-07-07 9:14 ` Sagi Grimberg
2015-07-07 9:14 ` Sagi Grimberg
2015-07-07 14:05 ` Steve Wise
2015-07-07 14:05 ` Steve Wise
2015-07-07 16:17 ` Jason Gunthorpe
2015-07-07 16:17 ` Jason Gunthorpe
2015-07-07 16:27 ` Sagi Grimberg
2015-07-07 16:27 ` Sagi Grimberg
2015-07-07 21:36 ` Jason Gunthorpe
2015-07-07 21:36 ` Jason Gunthorpe
2015-07-08 7:29 ` Sagi Grimberg
2015-07-08 8:13 ` 'Christoph Hellwig'
2015-07-08 8:13 ` 'Christoph Hellwig'
2015-07-08 10:05 ` Sagi Grimberg
2015-07-08 10:05 ` Sagi Grimberg
2015-07-08 10:20 ` 'Christoph Hellwig'
2015-07-08 10:20 ` 'Christoph Hellwig'
2015-07-08 11:08 ` Sagi Grimberg
2015-07-08 11:08 ` Sagi Grimberg
2015-07-08 17:14 ` Hefty, Sean
2015-07-08 17:14 ` Hefty, Sean
2015-07-09 8:46 ` Sagi Grimberg
2015-07-09 13:52 ` Chuck Lever
2015-07-10 19:34 ` Christoph Hellwig
2015-07-12 7:49 ` Sagi Grimberg
2015-07-13 16:50 ` Jason Gunthorpe
2015-07-14 8:06 ` Sagi Grimberg
2015-07-14 12:24 ` Tom Talpey
2015-07-14 12:24 ` Tom Talpey
2015-07-14 13:21 ` Sagi Grimberg
2015-07-14 13:21 ` Sagi Grimberg
2015-07-23 0:43 ` Hefty, Sean
2015-07-23 0:43 ` Hefty, Sean
2015-07-08 19:08 ` Jason Gunthorpe
2015-07-08 20:32 ` 'Christoph Hellwig'
2015-07-08 20:32 ` 'Christoph Hellwig'
2015-07-08 20:37 ` 'Christoph Hellwig'
2015-07-08 20:37 ` 'Christoph Hellwig'
2015-07-09 0:03 ` Jason Gunthorpe
2015-07-09 8:00 ` 'Christoph Hellwig'
2015-07-09 8:00 ` 'Christoph Hellwig'
2015-07-09 8:58 ` Sagi Grimberg
2015-07-09 8:58 ` Sagi Grimberg
2015-07-09 22:18 ` Doug Ledford
2015-07-09 22:18 ` Doug Ledford
2015-07-09 22:53 ` Jason Gunthorpe
2015-07-09 22:53 ` Jason Gunthorpe
2015-07-10 13:22 ` Tom Talpey
2015-07-10 16:11 ` Jason Gunthorpe
2015-07-10 16:11 ` Jason Gunthorpe
2015-07-10 17:56 ` Doug Ledford
2015-07-10 18:34 ` Chuck Lever
2015-07-10 18:42 ` Tom Talpey
2015-07-10 19:54 ` Jason Gunthorpe
2015-07-10 19:54 ` Jason Gunthorpe
2015-07-10 20:48 ` Jason Gunthorpe
2015-07-10 22:33 ` Doug Ledford
2015-07-10 22:33 ` Doug Ledford
2015-07-11 10:17 ` 'Christoph Hellwig'
2015-07-11 10:17 ` 'Christoph Hellwig'
2015-07-13 16:57 ` Jason Gunthorpe
2015-07-13 16:57 ` Jason Gunthorpe
2015-07-14 7:25 ` 'Christoph Hellwig'
2015-07-14 9:05 ` Sagi Grimberg
2015-07-14 15:35 ` 'Christoph Hellwig'
2015-07-14 17:26 ` Jason Gunthorpe
2015-07-15 7:10 ` Sagi Grimberg
2015-07-15 7:10 ` Sagi Grimberg
2015-07-10 22:30 ` Doug Ledford
2015-07-10 22:30 ` Doug Ledford
2015-07-10 20:57 ` Jason Gunthorpe
2015-07-10 22:27 ` Doug Ledford
2015-07-10 22:27 ` Doug Ledford
[not found] ` <20150710233417.GA8919@obsidianresearch.com>
2015-07-11 3:10 ` Doug Ledford
2015-07-11 3:10 ` Doug Ledford
2015-07-13 17:18 ` Jason Gunthorpe
2015-07-13 17:18 ` Jason Gunthorpe
2015-07-13 22:23 ` Tom Talpey
2015-07-11 16:37 ` Steve Wise
2015-07-12 10:46 ` Sagi Grimberg
2015-07-12 10:46 ` Sagi Grimberg
2015-07-14 19:25 ` Steve Wise
2015-07-14 19:25 ` Steve Wise
2015-07-14 19:29 ` Jason Gunthorpe
2015-07-14 19:32 ` Steve Wise
2015-07-14 19:32 ` Steve Wise
2015-07-14 19:37 ` Jason Gunthorpe
2015-07-14 19:55 ` 'Christoph Hellwig'
2015-07-14 19:55 ` 'Christoph Hellwig'
2015-07-14 20:10 ` Steve Wise
2015-07-14 20:10 ` Steve Wise
2015-07-14 20:29 ` Jason Gunthorpe
2015-07-14 20:29 ` Jason Gunthorpe
2015-07-14 20:40 ` Steve Wise
2015-07-14 20:40 ` Steve Wise
2015-07-14 20:44 ` Jason Gunthorpe
2015-07-14 20:44 ` Jason Gunthorpe
2015-07-14 20:54 ` Steve Wise
2015-07-14 20:54 ` Steve Wise
2015-07-14 20:59 ` Jason Gunthorpe
2015-07-14 20:59 ` Jason Gunthorpe
2015-07-14 20:50 ` Tom Talpey
2015-07-14 20:50 ` Tom Talpey
2015-07-15 6:50 ` 'Christoph Hellwig'
2015-07-15 19:12 ` Jason Gunthorpe
2015-07-15 19:12 ` Jason Gunthorpe
2015-07-16 6:41 ` Jason Gunthorpe
2015-07-16 6:41 ` Jason Gunthorpe
2015-07-16 8:04 ` 'Christoph Hellwig'
2015-07-16 8:04 ` 'Christoph Hellwig'
2015-07-16 16:13 ` Jason Gunthorpe
2015-07-16 16:13 ` Jason Gunthorpe
2015-07-15 8:47 ` Sagi Grimberg
2015-07-15 8:47 ` Sagi Grimberg
2015-07-15 12:19 ` 'Christoph Hellwig'
2015-07-15 12:19 ` 'Christoph Hellwig'
2015-07-15 19:17 ` Jason Gunthorpe
2015-07-15 19:17 ` Jason Gunthorpe
2015-07-14 20:46 ` Tom Talpey
2015-07-14 19:45 ` 'Christoph Hellwig'
2015-07-14 19:57 ` Jason Gunthorpe
2015-07-14 19:57 ` Jason Gunthorpe
2015-07-14 19:58 ` Steve Wise
2015-07-14 19:58 ` Steve Wise
2015-07-14 20:41 ` Jason Gunthorpe
2015-07-14 20:41 ` Jason Gunthorpe
2015-07-14 20:51 ` Steve Wise
2015-07-14 20:51 ` Steve Wise
2015-07-14 21:01 ` Steve Wise
2015-07-14 21:01 ` Steve Wise
2015-07-14 21:14 ` Jason Gunthorpe
2015-07-14 21:14 ` Jason Gunthorpe
2015-07-23 18:53 ` Hefty, Sean
2015-07-23 18:53 ` Hefty, Sean
2015-07-23 19:03 ` Steve Wise
2015-07-23 19:03 ` Steve Wise
2015-07-23 23:30 ` Hefty, Sean
2015-07-23 23:30 ` Hefty, Sean
2015-07-23 23:53 ` Jason Gunthorpe
2015-07-23 23:53 ` Jason Gunthorpe
2015-07-24 0:18 ` Hefty, Sean
2015-07-24 0:18 ` Hefty, Sean
2015-07-24 4:46 ` Jason Gunthorpe
2015-07-09 8:47 ` Sagi Grimberg
2015-07-09 8:47 ` Sagi Grimberg
2015-07-08 21:38 ` Tom Talpey
2015-07-08 23:36 ` Jason Gunthorpe
2015-07-09 11:02 ` Sagi Grimberg
2015-07-09 17:01 ` Jason Gunthorpe
2015-07-09 17:01 ` Jason Gunthorpe
2015-07-09 20:00 ` Tom Talpey
2015-07-09 21:16 ` Jason Gunthorpe
[not found] ` <20150709170142.GA21921-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2015-07-10 8:55 ` kernel memory registration (was: RDMA/core: Transport-independent access flags) Sagi Grimberg
[not found] ` <559F8881.7070308-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-07-10 16:35 ` Jason Gunthorpe
2015-07-11 10:31 ` 'Christoph Hellwig'
[not found] ` <20150711103153.GC14741-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2015-07-13 16:46 ` Jason Gunthorpe
[not found] ` <20150713164652.GC23832-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2015-07-14 8:24 ` kernel memory registration Sagi Grimberg
[not found] ` <55A4C73A.7080001-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-07-14 17:24 ` Jason Gunthorpe
2015-07-11 10:25 ` [PATCH V3 1/5] RDMA/core: Transport-independent access flags 'Christoph Hellwig'
2015-07-13 16:35 ` Jason Gunthorpe
2015-07-13 19:36 ` Tom Talpey
2015-07-13 20:15 ` Jason Gunthorpe
2015-07-14 9:10 ` Sagi Grimberg
2015-07-14 9:10 ` Sagi Grimberg
2015-07-14 15:36 ` 'Christoph Hellwig'
2015-07-14 15:36 ` 'Christoph Hellwig'
2015-07-14 15:47 ` Tom Talpey
2015-07-14 15:47 ` Tom Talpey
2015-07-14 16:22 ` Jason Gunthorpe
2015-07-14 16:22 ` Jason Gunthorpe
2015-07-14 7:37 ` 'Christoph Hellwig'
2015-07-14 9:22 ` Sagi Grimberg
2015-07-14 12:12 ` Tom Talpey
2015-07-14 12:12 ` Tom Talpey
2015-07-14 13:23 ` Sagi Grimberg
2015-07-14 14:45 ` Steve Wise
2015-07-14 14:45 ` Steve Wise
2015-07-14 15:40 ` 'Christoph Hellwig'
2015-07-14 15:40 ` 'Christoph Hellwig'
2015-07-08 8:11 ` 'Christoph Hellwig'
2015-07-06 7:58 ` Sagi Grimberg
2015-07-06 7:58 ` Sagi Grimberg
2015-07-06 14:39 ` Steve Wise
2015-07-06 14:39 ` Steve Wise
2015-07-05 23:22 ` [PATCH V3 2/5] RDMA/iser: Use transport independent MR allocation Steve Wise
2015-07-05 23:22 ` [PATCH V3 3/5] RDMA/isert: " Steve Wise
2015-07-05 23:22 ` Steve Wise
2015-07-05 23:22 ` [PATCH V3 4/5] svcrdma: " Steve Wise
2015-07-05 23:22 ` Steve Wise
2015-07-05 23:22 ` [PATCH V3 5/5] xprtrdma: " Steve Wise
2015-07-05 23:22 ` Steve Wise
2015-07-06 5:25 ` [PATCH V3 0/5] Transport-independent MRs Christoph Hellwig
2015-07-06 5:25 ` Christoph Hellwig
2015-07-06 14:24 ` Steve Wise
2015-07-06 14:24 ` Steve Wise
2015-07-07 9:01 ` 'Christoph Hellwig'
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='000c01d0b7f7$5c5f02c0$151d0840$@opengridcomputing.com' \
--to=swise@opengridcomputing.com \
--cc=bfields@fieldses.org \
--cc=dledford@redhat.com \
--cc=eli@mellanox.com \
--cc=hch@infradead.org \
--cc=linux-nfs@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=ogerlitz@mellanox.com \
--cc=roid@mellanox.com \
--cc=sagig@mellanox.com \
--cc=target-devel@vger.kernel.org \
--cc=trond.myklebust@primarydata.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.